การประมาณราคาเบื้องต้น สำหรับงานก่อสร้างอาคารด้วยวิธีเทคนิค โครงข่ายประสาทเทียม

Abstract

The budget estimation for construction works in the preliminary phase is standard and uses rough estimation. This method is quick and does not require any complicated steps but still lacks accuracy in price estimation. Currently, there is the development of computer technology and software that can be convenient for price estimation. One of the most widely used and effective techniques is neural network techniques. The building data is used for analysis, such as the number of floors, building area, or the number of passenger elevators. This research will apply neural network techniques to create a model for estimating the preliminary cost of building construction. It uses the data set of buildings that have already been built at Chiang Mai University. Therefore, it comes as a dataset for modeling. This research used 50 building data in Chiang Mai University to create a model with RapidMiner, a software used in modeling. The model accuracy is 90.46 percent, making it efficient to use as an alternative to preliminary price estimation to set up a construction budget.